Design of Tension Circular Flange Joints

in Tubular Structures
JUNJIE CAO and JEFFREY A. PACKER

INTRODUCTION
A s the use of circular steel tubes in different structures has
increased dramatically in recent years the connections be-
tween tubes have become very important. One of the most
important and common methods of making connections, es-
pecially field connections, in tubular structures is with bolted
circular flanges. The design of such a joint under an overall
axial tension force requires the determination of flange di-
mensions and the number and arrangement of bolts.
A typical circular flange joint is shown in Figure 1. Two
flanges are welded to the tubes and fastened by a number of
bolts and these bolts may be preloaded. Tis the tensile force
uniformly acting around the tubes. Under the action of T, the
flanges will bend and parts of the flanges may separate while
other parts remain in contact. Due to the existence of contact
force (prying force) between the two flanges, the total bolt
force is higher than the applied tension force. These factors
make the prediction of total bolt force in a joint and the
behavior of the joint more complex.
Because the joint in Figure 1 is symmetrical about the
middle surface between two flanges, only half of the joint
actually needs to be modelled in an analysis. The half joint is
very similar to the case when a circular column is connected
to a circular base plate under uplift loading, which may be the
governing load case for the base plate in single-storey build-
ings in some countries such as Australia.1 Hence the design
method for circular bolted flange joints in tubular structures
is also applicable to circular base plates connected to circular
columns under uplift loading.
Research work on bolted circular flange joints in tubular
structures has been carried out for many years and different
methods for the design of the joints have been developed.
From an investigation by Cao2 (1995) it has been found that
the methods currently used are either empirical rules or based
on some assumptions which are not applicable to all joints.
Further research work on circular flange joints, which in-
cludes finite element analyses, mathematical analyses and
tests, has been carried out and a better understanding of the
behavior of these joints has been obtained. In this paper
different methods currently used are compared and the defi-
ciencies and shortcomings of the methods are analyzed. A
new design method based on this research work is proposed
and design charts, which are in accordance with the AISC
LRFD specification for structural steel buildings3 (1993), are
created to simplify the design procedure, It will be seen that
the method is accurate, reliable and easy to use.
1. NOTATION
A Cross-sectional area of tube.
Ab Total nominal body area of bolts in a joint.
d Diameter of bolt.
D Diameter of tube.
P Diameter of bolt pitch circle.
Diameter of flange outside circle.
Nominal tensile strength of bolt.
Fyf
Yield stress of flange.
Yield stress of tube.
M Interactive bending moment between tube and
flange per unit length.
Maximum moment in a flange per unit length.
M
P
Plastic moment of flange plate per unit length.
Bending moment at tube/flange junction in a
flange per unit length.
M Bending moment at bolt pitch circle in a flange
per unit length.

Junjle Cao is research fellow, department of civil engineering,

University of Toronto.
Jeffrey A. Packer is professor, department of civil engineering,
University of Toronto.
Fig. 1. A typical circularflangejoint.

ENGINEERING JOURNAL / FIRST QUARTER / 1997 17

 My Yield moment of flange plate per unit length.
n Number of bolts in a joint.
P Total prying force in a joint.
r Radius of flange.
R Function of dimension and material property ra-
tios used for calculating Mmax.
S Function of dimension and material property ra-
tios used for calculating tf/t.
t Thickness of tube.
h Thickness of flange.
T Calculated value of total tensile force in a joint.
Tb Total bolt force in a joint.
Tb0 Total bolt preload in a joint.
% Total bolt tension capacity in a joint.
V Interactive radial force between tube and flange.
* Resistance factor of tube material.
H Slope of bolt force vs. tension force line.
Y Coefficient taking bolt preload into account in
calculation of Tb.
V Poisson's ratio.
C0 0 , CO l9
co2, co3,
co4 Coefficients for calculation of \i.
2. COMPARISON OF THE METHODS
CURRENTLY USED
Based on the test results for a large number of circular flange
joints some empirical rules for the design of these joints were
proposed by Rockey and Griffiths4 (1970) and other re-
searchers5,6 (1974, 1975) in the UK and these rules have been
incorporated in British Standard BS 81007'8 (1988, 1986) for
the design of towers and masts. In the Stelco design manual9
(1981) for hollow structural section connections, a method
for the determination of minimum flange thickness for a joint
was provided to ensure that the flange designed had sufficient
stiffness to resist prying action. This method was included in
an early design guide for hollow steel sections by CIDECT10
(Comite International pour le Developpement et 1'Etude de
la Construction Tubulaire) (1984). Tests and theoretical
analysis on circular flange joints have been carried out by
Igarashi et al.n,n (1985, 1987) in Japan and a method for the
determination of flange thickness and number of bolts in a
joint has been proposed. This method has been adopted in a
late design guide for circular hollow section (CHS) joints by
CIDECT13 (1991) and was also incorporated into a recent
book by Packer and Henderson14 (1992).
From research work by Cao2 and others, it has been found
that the bolt pitch circle in a joint should be as small as
possible and the flange outside circle is usually decided by
the arrangement of bolts. Hence the main task for the design
of a circular flange joint is to determine the thickness of flange
and the total bolt tension capacity required, which is used to
decide the diameter, number and grade of bolts, for the joint.
18 ENGINEERING JOURNAL / FIRST QUARTER /1997
A summary of the methods used by BS 8100, Stelco and
Igarashi et al. is presented in Table 1.
In order to compare the three different methods presented
in Table 1, they are used to determine the flange thicknesses
and the minimum numbers of bolts for the four joints listed
in Table 2. The flange bolt pitch circle diameters, flange
outside diameters and bolt diameters have been decided. The
yield stress for all the flanges is 36 ksi (A36), the yield stress
for all the tubes is 50 ksi (A50) and A325 bolts are used for
the joints. Three different levels of tension forces (7\, T2 and
T3 in Table 2) are applied to each of the joints, which corre-
spond to 90 percent, 75 percent and 60 percent of the squash
load (yield stress multiplied by cross-sectional area) of each
of the tubes. The design results, in terms of flange thickness
and number of bolts required, are presented in Table 3.
From Table 3 it can be seen that the results from the three
methods are quite different. The flange thicknesses from BS
8100 are much lower than those from Stelco. From Cao &
Bell's analysis15 (1996) it has been found that the total bolt
force in a joint may be much higher than 120 percent of the
tension force applied to the joint. Hence the flange thickness
and bolt capacity determined by BS 8100 for a joint are small
and a joint designed based on BS 8100 may not be safe,
especially for the cases with high tension forces. Because
Stelco's method attempts to reduce prying force in a joint to
zero, the flange designed by this method seems too thick.
Although thicker flanges for a joint can reduce prying force,
the prying force may not be eliminated totally so that the bolt
capacity may not be enough if no prying force is considered.
The method proposed by Igarashi et al. is based on the
analysis of plastic failures of flanges obtained using yield line
theory, but the yield line models were not proven by the tests
conducted by Igarashi et al. In some cases flanges may need
to be designed using elastic analysis, as recommended in EC
316 (clause 6.1.3). In addition, the bending moment between
tube and flange as well as the bolt stiffness are neglected in
Igarashi's analysis. Hence this method still cannot be used
confidently in all situations.
3. FURTHER RESEARCH WORK
Cao and Bell17 (1994) have carried out finite element analyses
using the package ABAQUS18 (1994) for circular flange
joints. Axisymmetric and 3-D models were used to analyze a
number of joints with different dimensions and bolt arrange-
ments. Bolt forces, contact forces between flanges, displace-
ments and stresses in the tubes and the flanges in these joints
were obtained from the analyses. The influences of bolt
preload, tension force and joint geometry on joint behavior
were investigated.
From the analyses it has been found that the discrete bolt
arrangement in a joint can be represented by a uniformly
annular arrangement without significant error. For most joints
analyzed with relatively high tension forces, only the edges
of the two flanges in a joint were in contact and this contact
 Table 1.
Summary of Design Methods
Method Details

BS8100
tf>2t'-*r- Tr>\.2T
Refs. 4-8 Fyf

, k

I * *
Stelco
hi

a DP/D
Refs. 9-10 Tr>T 0
* "l.O 1.1 1.2 1.3 1.4 1.5 1.6 1.7 l.B 1.9 2.0

Ratio of Bolt Pitch Circle Diameter (Dp)

to Tube Outside Diameter (D)

Igarashi s
etal.
" V^ V> h
[
-+
+
f3
1
f3HDf/Dp)
T

n
Refs. 11-14 f3+V/(S 4/(i k3 = 2 + *1
)3=^-L' " J /C1
- ^ -t

(prying) force was about 30 percent to 70 percent of the
applied tension force. For each joint analyzed the maximum
bending stress in the tubes was much higher than the maxi-
mum stress in the flanges. As the applied tension force was
increased the tubes at the junctions with the flanges were
found to yield first. The failure of a practical joint, in which
the flange thickness is considerably greater than the tube wall
thickness, was usually caused by excessive plastic deforma-
tions in the tubes and when a joint failed the flanges in the
joint either remained elastic or had just started to yield.
Based on the results from finite element analyses a model
for mathematical analysis for this type of joints has been
created by Cao and Bell15 (1996). In this model total bolt
force and prying force are modelled as uniformly annular
forces around the bolt pitch circle and the edge of the flange.
From the analysis of this model, formulas for the calculation
of total bolt force in a joint have been developed. Detailed
analysis about the prying action in circular flange joints has
also been done by Cao and Bell.19 The determination of total
bolt force in a joint is essential for the design of flanges and
bolts for the joint.
Further tests have been undertaken by Cao and Bell20
(1996) to confirm the results from finite element analyses and
mathematical analysis. In these tests static tension forces, bolt
forces, displacements and stresses have been measured and
the influences of bolt preloads and flange dimensions on the
behavior of flange joints have been investigated. Good agree-
ment between the test results and the results from finite
element analyses and mathematical analysis has been found.
4. DESIGN METHOD FOR CIRCULAR
FLANGE JOINTS
From the further research work15,17'19'20 described above the
behavior of circular flange joints has been investigated thor-
oughly. Although it is not practical to apply the formulas
derived in the work to regular designs directly, the formulas
really provide a basis for the formation of a rational design
method.
As explained before, the main task for the design of a
circular flange joint is to determine the thickness of flange
and the diameter, number and grade of bolts. Other flange
dimensions can be decided empirically first and then the
flange thickness can be determined based on the assessment
of maximum flange bending moment and the bolts can be
determined based on the assessment of total bolt force.
4.1 Determination of d, Dp and Df
Bolt diameter d should be decided first. Flange bolt pitch
circle diameter Dp should be kept as small as possible but the
clearance between the nut and the weld should be not less than
0.25 inches. The distance between flange outside radius and
bolt pitch circle radius can be taken as the same as the distance
between bolt pitch circle and tube surface, as suggested by
Packer and Henderson,14 i.e. Df- Dp-Dp- D. Cao2 has shown

ENGINEERING JOURNAL / FIRST QUARTER / 1997 19

 Table 2. Table 3.
Information for Joints Comparison of Different Methods

Tube Flange Bolt Tension Forces BS8100 Stelco Igarashi et al.

(inch) (inch) (inch) (kips) Tension
Joint Joint Force tf tf tf
No. Of No. (kips) (inch) ^min (inch) n
min (inch) rtmin

J-1 10 0.375 18 14 1.25 510 425 340 510 1.04 7.4 1.84 6.2 1.26 9.3

J-2 0.312 13 10 295 246 197 J-1 425 1.04 6.2 1.68 5.1 1.15 7.7

J-3 0.25 10 7.5 3

/4 168 140 112 340 1.04 4.9 1.51 4.1 1 03 6.2

J-4 0.203 % 80 67 54 295 0.87 6.7 1.43 5.6 0.99 8.4

J-2 246 0.87 5.6 1.31 4.6 0.90 7.0

Materials
197 0.87 4.5 1.17 3.7 0.81 5.6
Tubes: A50
168 0.69 6.8 1.11 5.6 0.79 8.5
Flanges: A36
Bolts: A325 J-3 1.02
140 0.69 5.6 4.7 0.72 7.0
JUr
112 0.69 4.5 0.91 3.8 0.64 5.6

that the increase of flange outer diameter in a joint has very 80 0.56 4.7 0.85 3.9 0.60 5.8
little influence on flange strength and can reduce total bolt J-4 67 0.56 3.9 0.77 3.2 0.55 4.8
force in the joint. Hence the method developed based on the
basis of D,f DP = DP D can actually be used for designs 54 0.56 3.1 0.69 2.6 0.49 3.9
with the case of D, P>DP D.

4.2 Calculation of Total Bolt Force 1 - C00CO!

(2)
Cao and Bell have carried out detailed analyses151719,20 for the co 4 - co 2 - C000>3
calculation of total bolt force in a circular flange joint. In co0, co1? co2 and co3 are non-dimensional coefficients related to
Figure 2 the "real curve" was obtained from a finite element four dimensional ratios of D/Df, D/Dp, t/D, and tf/1. co4 is
analysis for a typical joint with total bolt preload Tb0. It can be a coefficient, which represents the ratio of bolt axial stiffness
seen that the curve can be represented by two straight lines to flange bending stiffness at the bolt pitch circle. co4 relates
without significant error and the two lines can be expressed to the above four dimensional ratios and the ratio of the total
as: cross sectional area of the bolts, Ab, to the square of the flange
{Tb<\.\Tm) thickness tf. Hence JLI can be expressed as the following
USo
(1) general equation:
\iT + y-Th0 (Tb>\.\Tm)
tfA,
where:
n=/ D/D; D' t (3)

Tb - Total bolt force in the joint *s

Tbo = Total bolt preload in the joint
T = Tension force applied to the joint
In Equation 1, [i is the slope of the bolt force vs. tension force
line and y is used to take account of bolt preload. Formulas
for the calculation of |i and y have been derived by Cao and
Bell15 but they are tedious for practical designs. It has been
found15,1719'20 that the values of L
| L for some joints can be up to
1.7 but the values of y are usually small and can be taken as
0.1. Hence the evaluation of |Li is much more important in the
calculation of bolt force and for a joint under a relatively high
tension force, as shown in Figure 2, the effect of bolt preload
can be neglected. From Cao and Bell15 the formula for JLI can
be expressed as:
As a rational design for a joint, bolts should be decided
according to the total bolt force in the joint. In the AISC LRFD
design specification,3 the design tensile strength is based on
the lower value obtained from the limit states of yielding in
the gross section (gross area x yield strength x resistance
factor of 0.9) and fracture in the net section (net area x
ultimate strength x resistance factor of 0.75). For bolted
flange joints the tube net area equals the gross area. Further-
more, yielding of the gross section will typically be the design
criterion since, even for grade A50 hollow sections, the yield
to ultimate strength ratio = 50/62 = 0.81 < 0.75/0.90 = 0.83.
The design tension strength for all the bolts in a joint is
0.75A^ (AISC LRFD section J3.6) where Ft, the bolt "nomi-
nal tensile strength", is specified in the corresponding Table

20 ENGINEERING JOURNAL / FIRST QUARTER / 1997

J3.2. Hence the required bolt area, Abi to develop the maxi- flange simpler (but the moment M and force V have been
mum tension design capacity of the connected tube is: considered in the calculation of P and Tb). Then the flange can
be analyzed using the model shown in Figure 3(b), which is
0 9AF F a symmetrically loaded circular plate. Solutions for the cal-
A= - ^ ^0.75/?
^^=1.27tf(D-0^ (4)
* 0.75/? culation of bending and hoop moments at any position of the
So Equation 3 can be expressed as: flange plate can be obtained from an analysis carried out by
Cao and Bell22 (1992) (radius of central hole is zero here). If
r bolt preload is neglected, bolt force Tb and prying force P can
D D t U Fvi ^
(5) be expressed as Tb = \\Tand P = (\i-l)T. Hence the critical
[D/D;DW Ft
moments in a flange can be found to be:22
If Df- Dp = Dp- D, D / Df in Equation 5 can be deleted and Bending moment at the tube/flange junction Mrj:
JLL is a function of the other four ratios.
D2-(D-Q2-(Z^-D2)^
Mrjrj = ( l - v >
4.3 Calculation of Maximum Moment in a Flange 8TC D}
From the theoretical analysis21 (1993), finite element analy- R
ses17 (1994) and tests20 (1995) carried out by Cao and Bell, it + 2(l+v) VD-t - | i - In
D
(6)
has been concluded that the maximum moment in a blank
flange is either the bending moment at the tube/flange junc- Bending moment at the bolt pitch circle Mv:
tion or the bending moment at the bolt pitch circle.
,2A
As shown in Figure 3(a), the flange in a joint is acted on
v
T
(1-v) %z
Dj V-
(D-t)
by a uniformly annular tension force T, bolt force Tb, prying 8JI P
V J
force P(P = Tb -T), interactive moment M and interactive
Df
radial force V between the flange and tube. Because the tube + 2(l+v)ln^0i-l) (7)
is relatively thin, the interactive moment M is low and the
stress caused by this moment is usually much lower than the It is noted that the existence of prying force in a joint will
stresses caused by Tb and P. The effect of radial force Vis also reduce the moments at the tube/flange junction of the flange.
small so that the moment M and force V in Figure 3(a) can be Thicker flanges are usually needed for a joint if prying force
neglected to make the calculation of maximum moment in a is neglected in the assessment of flange strength.
If v = 0.3 is used and Equation 5 is considered, the maxi-
mum moment, Mmax, for a flange can be obtained from Equa-
400 ~ i i i i r
tions 6-7 and expressed as:
f A
n r + y7V D D t ^ DDttf Fvl
M^=TR =T R (8)
D; D; D' ^ , Df'Dp9D't* Ft
v f p J
300

-
O
PL, 200

O
PQ M M v
,_
$ T
L c^ad ^
o b0
H 100

100 200 300

(b)
Tension Force T (kN)

Fig. 2. Relationship between bolt force and tension force. Fig. 3. Calculation of bending moment in flange.

ENGINEERING JOURNAL / FIRST QUARTER / 1997 21

Similar to |i in Equation 5, if Df - Dp = Dp - D is used, the D/Dp and t/D. Although it is not easy to find an explicit
maximum bending moment in the flange in Equation 8, solution for Equation 12 the relationship between tf/t,
Mmax, is the product of tension force T and a function of four D/Dp and t/D, for a material combination, can be deter-
ratios (without D/Df in function R). mined by using an iteration process.
The preferred yield strength for round Hollow Structural
4.4 Design Charts for Flange Thickness and Bolt Sections in the U.S. is 50 ksi (A50), the yield stress for
Capacity structural plate is generally 36 ksi (A36) and the nominal
If a flange is designed elastically the maximum bending strengths for bolts are 90 ksi (A325) and 113 ksi (A490). The
moment, Mmax, in the flange should be limited not to exceed results from Equation 12 for these two types of bolts are
the yield moment, My, of the flange: presented in Figures 4 and 5 (solid lines).
Equation 5 shows that after the material properties Fyt and
M <M =Ft t. (9) /^are decided the value of \i is a function of three dimensional
ratios (D/Dp, t/D and tf/t). If the critical values of tf/t,
Inserting Equation 8 gives: determined by Equation 12 are used for Equation 5, L ]L is also
a function of D/Dp and t/D. The results for \i (the prying
6 T ratio) are also presented in Figures 4 and 5 (dashed lines) for
(10)
T**** the two types of bolts. As explained before, the effect of bolt
preload is negligible if tension force is relatively high, so the
If the joint is designed to develop the full tension capacity of
value of ja represents the ratio of total bolt force to applied
the tube, T= 0.9AFVP then Equation 10 can be expressed as:
tension force.
D-tR, Similarly, if a flange is designed plastically using
> 6(0.9)n ; R (11) Mp(Mp = 1.5M>(), of the flange in Equation 9, the charts for
t2 Fyf
flange plastic design can be also obtained, as shown in Figures
Considering Equation 8 andD f -D p = Dp-D and taking the 6 and 7.
critical state for the flange gives: It can be seen that after the tube and flange pitch circle
diameter for a joint are specified, the ratio t/D (vertical axis

w D tF
t byf D; D't9 F,

Hence the ratio of flange thickness to tube thickness, tf/t, is

' FJ
(12)
a function of two dimensional ratios and two material prop-
erty ratios. If materials for tube, flange and bolt are deter-
mined the ratio tf/t is a function of two dimensional ratios
in Figures 4 to 7) and the ratio D/Dp (horizontal axis in
Figures 4 to 7) can be determined. Then the critical value for
ratio tf/t and the corresponding ratio of Tb / J e a n be deter-
mined from the figures easily for the joint.
Although the charts are designed to develop the full tension
capacity of the tube (0.9AFyt) at a joint, the charts can also be

Flange Elastic Design Flange Elastic Design

0.4 0.5 0.6 0.7 0.8 0.9 0.4 0.5 0.6 0.7 0.8 0.9
Ratio of Tube Diameter to Bolt Pitch Circle Diameter {DID ) Ratio of Tube Diameter to Bolt Pitch Circle Diameter (DID )

Fig. 4. Design chart for flange elastic design Fig. 5. Design chart for flange elastic design
using A3 2 5 bolts for circular flange joints. using A490 bolts for circular flange joints.

22 ENGINEERING JOURNAL / FIRST QUARTER / 1997

used for joints with lower tension loads. If the tension load elastic and plastic designs for the joints listed in Table 2. The
for a joint is x A /, the ratio of required flange thickness to bolt pitch circles of the flanges have been determined based
tube thickness for the joint can be determined as following: on the bolt diameters selected, as presented in Table 2, so the
ratios of t/D and D/Dp can be obtained. Then the ratios of
l
tf/1 and Tb / 7 can be found from Figure 4 and Figure 6 (using
f
A325 bolts), for both flange elastic design and flange plastic
V) X
design respectively. Three different levels of tension force,
representing 90 percent, 75 percent and 60 percent of the
where:
squash load, are considered for each of the joints. The flange
(A thickness and the minimum number of bolts for each of the
- = ratio of flange thickness to tube thickness for the joints under tension force Tx can be found directly from the
v )x joint ratios of tf/1 and Tb/ T. The flange thicknesses for each of the
(tf) joints under tension forces T2 and % can be obtained by
- = ratio of flange thickness to tube thickness multiplying the flange thickness under Tx by V0.75/0.9 and
v / 9 obtained from Figures 4 to 7 VO.6/0.9, respectively. The minimum number of bolts, for
If a material other than A50 is used for the tube or a material each of the joints, under tension forces T2 and T3 can be
other than A36 is used for the flange for a joint, the ratio of obtained by multiplying the minimum number under Tx by
tf/1 can be obtained by multiplying the ratio from Figures 4 0.75 /0.9 and 0.6/0.9, respectively.
to 7 by The number of bolts for a joint should be not less than four
and the circumferential distance between two bolts should be
greater than 3d, as required in AISC LRFD design specifica-
tion.3 If the number of bolts for a joint is too big or too small
V/vX50
the diameter of bolts should be changed and the joint should
be redesigned.
where
It can be seen, from Table 4, that although the flange
Fyt and Fyf = yield stresses of the tube and the flange in ksi thicknesses are reduced by using plastic design, the numbers
of bolts are increased. From Cao and Bell's tests20 it has been
The ratio of total bolt force to applied tension force can be found that the deformation of flanges can cause bending of
taken to be the same. bolts so that flanges should be rigid enough to resist excessive
deformations. Hence the charts for flange elastic design may
5. EXAMPLES be preferable for some situations.
The design charts, Figures 4 to 7, can be applied to the designs Comparison of Table 3 and Table 4 shows that the flange
of circular flange joints easily. Table 4 shows the results for

Flange Plastic Design Flange Plastic Design

0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Ratio of Tube Diameter to Bolt Pitch Circle Diameter (D/Dp) Ratio of Tube Diameter to Bolt Pitch Circle Diameter (D/Dp)

Fig. 6. Design chart for flange plastic design Fig. 7. Design chart for flange plastic design
using A3 2 5 bolts for circular flange joints. using A490 bolts for circular flange joints.

ENGINEERING JOURNAL / FIRST QUARTER / 1997 23

 Table 4.
Examples

Tension Forces
Bolt
Dia. Parameters Tx (90%) T2 (75%) T3 (60%)

Joint Do tf tf tf
No. inch t/D D/Dp tf/t Tb/T inch inch inch
^min rtmin rtmin

Flange Elastic Design Using Figure 4

J-1 1.25 0.0375 0.714 4.09 1.59 1.53 9.8 1.40 8.1 1.25 6.5

J-2 1 0.0446 0.700 3.96 1.51 1.24 8.4 1.13 7.0 1.01 5.6

J-3 % 0.0500 0.667 4.09 1.38 1.02 7.8 0.93 6.5 0.84 5.2

J-4 % 0.0677 0.600 4.54 1.10 0.92 4.2 0.84 3.6 0.75 2.9

Flange Plastic Design Using Figure 6

J-1 1.25 0.0375 0.714 3.51 1.67 1.32 10.3 1.20 8.6 1.08 6.9

J-2 1 0.0446 0.700 3.24 1.65 1.01 9.2 0.92 7.6 0.83 6.1

J-3 % 0.0500 0.667 3.16 1.63 0.79 9.2 0.72 7.6 0.65 6.1

J-4 % 0.0677 0.600 3.05 1.55 0.62 6.0 0.57 5.0 0.51 4.0

thicknesses from elastic design in Table 4 are between the
results from Stelco and the results from Igarashi et al. The
flange thickness results from Igarashi et al. are very similar
to the results from plastic design in Table 4, so the flanges
designed in accordance with Igarashi et al. are rather flexible.
Also, the plastic design method presented herein is a little
more conservative than Igarashi et al. with regard to the
requirements for the number of bolts.
America and the American Institute of Steel Construction will
shortly be publishing new section properties for hollow sec-
tions produced in accordance with ASTM A500. These will
be based on a tube wall thickness, for design purposes, of 0.93
times the nominal tube wall thickness. The design procedures
and charts presented herein will still be applicable for these
new properties as the tube wall thickness is an explicit vari-
able in Figures 4 to 7.

5. CONCLUSIONS
Based on previous theoretical research work and tests on
circular flange joints in tubular structures, a new design
method for these joints is proposed. Design charts, which are
in accordance with the AISC LRFD design specification,
have been produced and can be used to decide flange thick-
ness and number of bolts for a joint very easily. These design
charts are produced for one tube grade and one flange grade
but can be modified for various grades of tube and flange plate
materials, both A325 and A490 bolts, and various levels of
tension load applied to the connection. It has been shown that
the method is accurate, reliable and easy to use.
The design examples given in the paper use the nominal
geometric properties of several ASTM A500 grade round
Hollow Structural Sections. However, the ASTM A500 stand-
ard23 permits a hollow section wall thickness as much as 10
percent below the nominal wall thickness, without specifying
any mass (or weight or cross-sectional area) tolerance. This
can have a major negative effect on the assumed (nominal)
structural properties24 as most U.S. manufactures now tend to
produce thin sections. Both the Steel Tube Institute of North
REFERENCES
1. Packer, J. A., "Design with Structural Steel Hollow Sec-
tions," Australia Institute of Steel Construction Seminar
Notes, Australia, March, 1996.
2. Cao, J. J., 'Tension Circular Flange Joints in Tubular
Structures," Ph.D Thesis, Department of Civil and Struc-
tural Engineering, UMIST, UK, 1995.
3. American Institute of Steel Construction, Load and Re-
sistance Factor Design Specification for Structural Steel
Buildings, Chicago, IL, 1993.
4. Rockey, K. C , and Griffiths, D. W., The Behaviour of
Bolted Flanged Joints in Tension, Conference on Joints in
Structures, University of Sheffield, England, 1970.
5. British Steel Corporation Tubes Division, Bolted Tension
Flange Joint Research (Ring Flanges Joining CHS), Re-
search Report CE 71/46, Corby, UK, 1974.
6. British Steel Corporation Tubes Division, Bolted Tension
Flange Joint Research (Blank Circular Flanges Joining
CHS), Research Report CE 71/46, Corby, UK, 1975.
7. British Standards Institution, BS 8100 Lattice Towers and
Masts, London, UK, 1988.

24 ENGINEERING JOURNAL / FIRST QUARTER / 1997

 8. British Standards Institution, DD 133 Draft for Develop-
ment Code of Practice for Strength Assessment of Mem-
bers of Lattice Towers and Masts, London, UK, 1986.
9. Stelco, Inc., Hollow Structural Sections Design Manual
for Connections, 2nd Ed., Hamilton, Canada, 1981.
10. CIDECT, Construction with Hollow Steel Sections, Brit-
ish Steel Corporation, Corby, UK, 1984.
11. Igarashi, S., Wakiyama, K., Inoue, K., Matsumoto, T., and
Murase, Y., "Limit Design of High Strength Bolted Tube
Flange Joints (Part 1)," Trans. ofA.I.J., No.354,1985, pp.
52-66.
12. Igarashi, S., Inoue, K., and Matsumoto, T., "Limit Design
of Bolted Circular Tube Flange Joint Subjected to Ten-
sion," Proc of the Int. Conference on Steel and Aluminium
Structures, Cardiff, UK, 1987, pp. 825-834.
13. Wardenier, J., Kurobane,Y., Packer, J. A., Dutta, D., and
Yeomans, N., Design Guide for Circular Hollow Section
(CHS) Joints under Predominantly Static Loading,
CIDECT (ed) and Verlag TUV Rheinland GmbH, Koln,
Germany, 1991.
14. Packer, J. A., and Henderson, J. E., Design Guide for
Hollow Structural Section Connections, Canadian Insti-
tute of Steel Construction, Toronto, Canada, 1992.
15. Cao, J. J., and Bell, A. J., "Determination of Bolt Forces
in a Circular Flange Joint under Tension Force," Int. J.
Pres. Ves. & Piping, Vol. 68, No. 1, 1996, pp. 63-71.
16. European Committee for Standardization, Eurocode 3:
Design of Steel StructuresPart 1.1: General Rules and
Rules for Buildings, ENV 1993-1-1:1992E, London, UK,
1992.
17. Cao, J. J. and Bell, A. J., "Finite Element Analysis of
Circular Flange Joints under Tension Forces," Proc. of
Ninth UKABAQUS User Group Conference, Exeter Col-
lege, Oxford, UK, 1994.
18.Hibbit, Karlsson & Sorensen, Inc., ABAQUS/Standard
Users Manual Version 5.4, USA, 1994.
19. Cao, J. J., and Bell, A. J., "Prying Action in Bolted
Circular Flange Joints in Tubular Structures," Proc. of the
7th Int. Symposium on Tubular Structures, Miskolc, Hun-
gary, 1996.
20. Cao, J. J. and Bell, A. J., "Experimental Study of Circular
Flange Joints under Tension Forces", J. Strain Analysis,
Vol. 31, No. 4, 1996, pp. 259-267.
21. Cao, J. J., and Bell, A. J., "Elastic Analysis of a Circular
Flat Flange Joint Subjected to Axial Force," Int. J. Pres.
Ves. & Piping, Vol. 55, No. 3, 1993, pp. 435^49.
22. Cao, J. J., and Bell, A. J., "General Solutions for a Circular
Flat Plate with a Central Hole Loaded by Transverse
Force or Bending Moment Uniformly Distributed along
a Circumference", Int. J. Pres. Ves. & Piping, Vol. 51, No.
2, 1992, pp. 155-173.
23. American Society for Testing and Materials, Standard
Specification for Cold-Formed Welded and Seamless
Carbon Steel Structural Tubing in Rounds and Shapes,
ASTM A500-93, ASTM, Philadelphia, PA, 1993.
24. Packer, J. A., "Overview of Current International Design
Guidance on Hollow Structural Section Connections",
Proc. of the 3rd. International Offshore and Polar Engi-
neering Conference, Singapore, Vol. IV, 1993, pp. 1-7.
ENGINEERING JOURNAL / FIRST QUARTER / 1997 25